1. Charm and J/PSI measurement at RHIC Y. Akiba (KEK) August 9, 2002 Riken Summer study 2002

2. Outline PHENIX experiment Single electron measurement –Charm/botom measurement via single lepton –Run-1 (130 GeV) results (PRL…) –Run-2 (200 GeV) prelimnary result J/PSI measurement at RHIC –pp  J/PSI  , ee –Au+Au  J/PSI  ee

3. PHENIX collaboration

4. PHENIX Detector Electrons ( |  | < 0.35) Charged tracks (Beam-Beam, Drift Chamber, Pad Chambers) + RICH rings + EM Calorimeter clusters Muons (1.2 <  < 2.2) Muon Identifier roads + Muon Tracker tracks

5. Charm measurement Direct method: Reconstruction of D-meson (e.g. D 0  K  ). Very challenging without measurement of displaced vertex ++ Indirect method: Measure leptons from semi- leptonic decay of charm. This method is used by PHENIX at RHIC

6. History single electron at the ISR Single electron was observed at the ISR in early 1970 ’ s, before charm discovery. –F. W. Busser et al,PLB53,212 –F.W.Busser et al,NPB113,189 It has been later interpreted as a signal of open charm. –I.Hunchlife and C.H.Llewellyn Smith, PLB61,472 –M. Bourquin and J.-M.Gaillard, NPB114,334 Signal level is e/  ~1-2x10 -4 at the ISR Signal level at RHIC is higher (e/  ~3x10 -4 in p+p) because of higher energy.

7. single electrons at high pt At RHIC (s NN ½ ~200GeV), the electron signal from charm is expected at e/  ~ 3 x10 -4 in p+p The e/  ratio is higher ( ~10 -3 ) in Au+Au collision –Production of the high pt pions is strongly suppressed relative to binary scaling –Production of charm quark roughly scale with binary collisions. B decay contribution can be dominant in pt>4 GeV/c (Simulation)

8. Single electron in RUN-1(130GeV) Fully corrected single electron spectra in PHENIX. The spectra includes background such as Dalitz decays and photon conversions.

9. Sources of electrons Dalitz decays of  0, ,  ’, ,   0  ee ,   ee , etc) Di-electron decays of , ,  Photon conversions Kaon decays Most of the background are PHOTONIC Charm decays Bottom decays Other –Thermal di-leptons –Conversion of direct photons Background Signal

10. Background subtraction (cocktail method) The measured electron spectra includes trivial background from light hadron decays such as  0 Dalitz decay and photon conversions. The background is estimated using a hadron decay generator that is constrained by pion measurement by PHENIX  0  e + e -   0   e+e-e+e- ~80% of background Proportional to pion conversion ~ 1.9 x Dalitz in PHENIX   e + e -     e+e-e+e- ~20% of  0 contribution at high pt Other contributions: small

11. Data/background ratio in RUN-1 The Upper panel shows data/background ratio as function of p t for min. bias collisions. The data show excess above background in pt>0.6 GeV/c. Central collision data also show similar excess. Peripheral data do not have enough statistics The low panel shows the relative contribution to the background from various sources.

12. Background-subtracted single electron spectra Spectra of single electron signal is compared with the calculated charm contribution. Charm contribution calculated as EdN e /dp 3 = T AA Ed  /dp 3 –T AA : nuclear overlap integral –Ed  /dp 3 : electron spectrum from charm decay calculated using PYTHIA The agreement is reasonably good.

13. Charm cross section from the electron data  We can estimate the charm yield by assuming that all single electrons above the background are from charm  Neglect other possible sources such as thermal  and di-leptons  Charm yield can be over-estimated.  By fitting the PYTHIA electron spectrum to the data for pt>0.8 GeV/c, we obtained charm yield Ncc per event.  The charm cross section per binary NN collision is obtained as T AA is nuclear overlap integral ~ NN integrated luminosity per event T AA (0-10%)=22.6±1.6/mb T AA (0-92%)=6.2±0.4/mb  Charm yield per binary collision

14. Comparison with other experiments PHENIX single electron cross section is compared with the ISR data Charm cross section derived from the electron data is compared with fixed target charm data Solid curves: PYTHIA Shaded band: NLO QCD Assuming binary scaling, PHENIX data are consistent with  s systematics (within large uncertainties)! Expect increase of cross section by factor ~2 going from 130 to 200 GeV

15. Photon converter method in RUN2 Most of the background electrons are photonic. They are internal or expernal conversion of photons The signal electron from charm/bottom decays are non-photonic. In Run 2 we had a special run where we placed a photon converter near the interaction point.The converter increases the yield of photonic electrons by a fixed factor.  We can separate photonic electron (background) and non-photonic electron (charm/bottom signal) by comparing the electron spectrum in the converter run and the normal run. Method has many advantages and completely independent systematic errors from previous “cocktail” subtraction method. γ e+ e- Converter Au

16. RUN-2 (200GeV) electron spectra When a large fraction of the electrons come from non-photonic sources, the converter and non-converter data will approach each other. This is observed at high pT.

17. RUN2 single electron result The yield of non-photonic electron at 200 GeV is higher than 130 GeV The increase is consistent with PYTHIA charm calculation (  cc (130GeV)=330  b   cc (200GeV)=650  b) Large systematic uncertainty due to material thickness without converter. The error will be reduced in the final result.

18. Centrality Dependence

19. Observations Our data is consistent with binary scaling within our current statistical and systematic errors. NA50 has inferred a factor of ~3 charm enhancement at lower energy. We do not see this large effect at RHIC. PHENIX observes a factor of ~3-5 suppression in high p T  0 relative to binary scaling. We do not see this large effect in the single electrons. Initial state high pt suppression excluded? smaller energy loss for heavy quark ? (dead cone effect) NA50 - Eur. Phys. Jour. C14, 443 (2000). N part Enhancement of Open Charm Yield Binary Scaling PHENIX Preliminary

20. Outlook for single electron measurement Refine converter subtraction analysis –Reduce systematic uncertainties at low pt Cocktail subtraction method –Systematic check of converter method result –Measurement at high pt with full RUN-2 statistics Reference measurement in p+p Expected results from RUN-2 –Total charm yield  shadowing effect at x~0.01 –Pt distribution of charm  energy loss for charm quark? –Centrality dependence –B quark contribution? (should become dominant in pt>4GeV/c)

21. J/  Physics J/  production in N+N –J/  production mechanism –Spin dependence J/  production in Au+Au –Suppression by QGP formation? –Enhancement by thermal production? PHENIX can measure J/   ee in |y|<0.35 and J/PSI   in 1.2<|y|<2.4 –Wide kinematics coverage –Independent measurement in two channels PHENIX can measure of both open charm and J/  –VERY strong constraint on the models of J/  production

22. J/   ee in p-p N J/  = 24 + 6 (stat) + 4 (sys) Bd  /dy = 52 + 13 (stat) + 18 (sys) nb J/PSI measurement in mid-rapidity (|y|<0.35) All triggers.

23. J/    in pp 1.2 < y < 1.7 N J/  = 26 + 6 + 2.6 (sys) B d  /dy = 49 + 22% + 29% (sys) nb 1.7 < y < 2.2 N J/  = 10 + 4 + 1.0 (sys) B d  /dy = 23 + 37% + 29% (sys) nb

24. J/    pt distribution

25. J/  Bd  /dy in p+p(200GeV)

26. Comparison with lower energy PHENIX preliminary CEM predictions (Phys.Lett.B390:323-328,1997)

27. J/   e + e - in Gold-Gold ! N=10.8  3.2 (stat)  3.8 (sys)N=5.9 +  2.4 (stat)  0.7 (sys)

28. N=4.5  2.1 (stat)  0.5 (sys) N=3.5  1.9 (stat)  0.5 (sys) J/   e + e - in Gold-Gold !

29. J/  B-dN/dy (Au-Au and p-p) B ee dN/dy 0-90% (M.B.) 1.4±0.4±0.6×10 -4 0-20% 3.9±1.6±1.0×10 -4 20-40% 2.5±1.2±0.7×10 -4 40-90% 6.9±3.7±1.9×10 -5

30. J/  B-dN/dy per binary collision

31. Model comparison (1) Binary scaling The statistical confidence level is defined as the probability, based on a set of measurements, that the actual probability of an event is better than some specified level.

32. Model comparison (2) Nuclear absorption * We hope to measure the nuclear absorption in d-A at RHIC energies in the next two years. See poster by D.Silvermyr.

33. Model comparison (3) NA50 data

34. Model comparion (4) (J/  )/N cc re-generation models Some models of J/  formation predict that J/Y yield per Ncc is higher in Au+Au than in pp PHENIX J/  and single electron data indicates that N J/  /N cc ~ 3 (± ?) ×10 -3 at mid-rapidity  Models with large J/  formation are exclude? L. Grandchamp and R. Rapp, hep-ph/0205305 R. L Thews, hep-ph/0206179

35. Model comparison Present data can not discriminates between –Binary scaling –Normal nuclear suppression –Suppression pattern seen by NA50 due to large statistical and systematical unceratinties. Some models predict J/  yield that is much larger than the binary scaling model. They are at the edge of acceptance. Expect factor ~2 increase of statistics in both p+p and Au+Au using all dataset in RUN-2 Combined with open charm data, the RUN-2 J/PSI data can give a constraint on some of J/  models.

36. Summary PHENIX has measured single electrons from charm decay in Au+Au collision at RHIC Single electron data is consistent with the binary scaling –No or little nuclear shadowing effect in charm produciton? –No or little energy loss for charm quark? –Important input to understand high pt suppression –We are working to reduce systematic errors of the measurement PHENIX made the first measurement of J/  in p+p and Au+Au at RHIC –Present statistics is too limited to draw any conclusion on J/PSI suppression, but some J/  formation models can be excluded from the data –We have about factor 2 more statistics in RUN-2 –Future high statistics Au+Au run is needed to make a decisive measurement of J/  suppression at RHIC

37. BACKUPS

38. Background:  0 Dalitz Background from Dalitz decays and di-electron decays are calculated using a hadron decay generator. The  0 spectrum of the generator is determined by combined fit to PHENIX  0 and  ± data. The systematic in the spectrum is estimated by changing the input  ± and  0 spectrum within their systematic errors.

39. Background: other hadrons Pt distribution of other hadrons are obtained from pi0 spectrum by m T scaling p t  sqrt(p t 2 +m h 2 -m  2 ) Spectra shapes of K ±,p ± agree with this scaling within 20% Particle ratios at high pt: –  = 0.55 –  ’  = 0.25 –  = 1.0 –  = 0.40 Assign conservative systematic error of 50% to these ratios Data/Mt scaled spectrum (p) data/Mt scaled spectrum (K + )

40. PYTHIA and charm data Use PYTHIA 6.152 Modified parameters – Mc = 1.25 GeV – K = 3.5 – = 1.5 GeV/c Charm cross section by fixed target experiments are reasonably reproduced. Pt distribution of D-mesons by E791 and BEATRICE are also reproduced. s 1/2 (GeV) p T 2 (GeV 2 /c 4 )  cc in pN,  N d  /dp T 2

41. PYTHIA and ISR single electron Electron spectrum from charm decay calculated with the tuned PYTHIA is compared with the single electron data at the ISR. The calculation reasonably reproduces the single electron data. p T (GeV/c)